Re:Constitutional AI: RLHF on Steroids
by eterevsky on Tuesday May 09, 2023 @04:47AM
(#35871669)
"Accurate" is a component in "helpful".
[ Parent
| on HN ]
Indeed. It could also be a component in "ethical," depending on what your ethics are.
No it isn't. Harmlessness, helpfulness, honesty are separate components. See paper. A model could be very helpful and tell you everything you want to hear, but at a cost of engaging in deception. E.g. in regards to politically incorrect information.